lkml.org 
[lkml]   [2022]   [Nov]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] 9p: fix crash when transaction killed
Schspa Shi wrote on Wed, Nov 30, 2022 at 09:15:12PM +0800:
> >> If the req was newly alloced(It was at a new page), refcount maybe not
> >> 0, there will be problem in this case. It seems we can't relay on this.
> >>
> >> We need to set the refcount to zero before add it to idr in p9_tag_alloc.
> >
> > Hmm, if it's reused then it's zero by definition, but if it's a new
> > allocation (uninitialized) then anything goes; that lookup could find
> > and increase it before the refcount_set, and we'd have an off by one
> > leading to use after free. Good catch!
> >
> > Initializing it to zero will lead to the client busy-looping until after
> > the refcount is properly set, which should work.
>
> Why? It looks no different from the previous process here. Initializing
> it to zero should makes no difference.

I do not understand this remark.
If this is a freed request it will be zero, because we freed the request
as the refcount hit zero, but if it's a newly allocated request then the
memory is uninitalized, and the lookup can get anything.

In that case we want refcount to be zero to have the check in
p9_tag_lookup to not use the request until we set the refcount to 2.


> > Setting refcount early might have us use an re-used req before the tag
> > has been changed so that one cannot move.
> >
> > Could you test with just that changed if syzbot still reproduces this
> > bug? (perhaps add a comment if you send this)
> >
>
> I have upload a new v2 change for this. But I can't easily reproduce
> this problem.

Ah, I read that v2 as you actually ran some tests with this, sorry for
the misuderstanding.

Well, it's a fix anyway, so it cannot hurt to apply...
--
Dominique

\
 
 \ /
  Last update: 2022-11-30 14:35    [W:0.066 / U:0.072 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site